Sending a document is not the end of it — knowing which section someone lingered on, and whether they are reading it right now, is what turns a follow-up into a guess or a fact.
A visit is a session with a start and an end, and reading time counts only while the document is on screen. Sessions abandoned mid-read are closed out and counted for what they were.
proposal.quix was opened
2 minutes ago · 4m 18s
scope-of-work.quix — 2nd visit
Yesterday · 6m 52s
Daily digest · 3 documents read
Today, 08:00 your time
An email and an in-app notification follow a real visit within a minute or two. Bots, your own opens and glances too short to mean anything never reach you, and a reader who keeps refreshing is not three notifications.
Open sessions show as live for as long as they last, so you can tell when the document is in front of someone.
Ask an unknown reader for an address before the document opens. It is typed in by the reader and never verified, and the product labels it that way everywhere it appears.
Download a document's visit history as a PDF — a dated record you can put in a project file or send to the client.
Every heading in a shared document is timed on its own. Not just that the document was opened — which part of it was actually on screen, in the document’s own order, for as long as that lasted.
The section that took most of the time.
Named and timed against the total — “most of their attention went to Pricing” is a sentence the page writes for you, not a chart you have to read. On a proposal it is usually the price, and knowing that changes what you say when you follow up.
Reached, but never held them.
On screen for a moment and then gone. A section that gets to this stage and a section a reader never opened the document far enough to see are different findings.
Past the point where the reading stopped.
Nothing to hold or pass over here — the reader closed the document before reaching it, which is reported as its own finding rather than folded into “passed over”.
A held section, visited a second time.
Returning to a section that once held them is the clearest signal here — it is what deciding about something looks like from outside.
Measured only while the tab was in front of the reader and they were doing something in it — scrolling, moving a cursor. That is presence on screen, not comprehension, and the page never says otherwise. Sections are the document’s own headings, and a single shared document is what carries them.
Country, city, device, browser and operating system, tallied from the same sessions the reader list already shows — one entry per person, so a reader who came back four times is one line here too, not four.
Country
City — approximate
Device
About 1 in 4 city guesses is wrong: an address often resolves to the reader’s internet provider, not their town.
A document goes to your own team before it goes out, often down the same link. Every session is checked against who is signed into the workspace that owns the share, and a match is counted on its own — kept out of the view count, the unique-people figure, the reading-time average, the digest, the view alert and the exported report. All of those answer for clients only.
3 reads
by 2 people inside your workspace
One link often reaches more than one person. Choosing a reader, a country or a city narrows the counters, the chart, the reader list, the audience breakdown, the section attention card and the exported report together, so every figure on the page is answering about the same person at once.
A name is only worth what produced it. Every reader lands in one of three tiers, and the tier is shown next to the name every time it appears.
They were signed in when they opened it.
The address comes from the Quixli account that opened the link, not from anything typed on the way in. This is the only tier that is proof of who opened the document.
They typed an address into the email gate.
Evidence of intent — somebody wanted the document enough to hand over an address — and not evidence of identity, because nobody checked it. It carries a self-reported badge everywhere it is shown, including in the alert email.
Nothing about the reader is established.
If you mailed the link to a named address, that address is shown as “someone with the link sent to …” rather than as a name. A link can be forwarded, so the address is a fact about your mailing list, not about your reader.
The label and the tier are resolved together, from the same record, so a name can never be shown at a confidence it did not earn.
Quixli measures what was on the reader’s screen and for how long. It cannot measure what they took in, and it does not guess.
| Question | With a Quixli link | With a file or plain URL |
|---|---|---|
| Was it opened at all? | Yes, with a timestamp | Unknown |
| How long did they read? | Measured per session | Unknown |
| Did they come back? | Every visit, listed | Unknown |
| Are they reading now? | Live, while it lasts | Unknown |
| Which parts held them? | Per section, on screen | Unknown |
| Who was it? | If named or self-reported | Unknown |
| Did they take it in? | Not measured, by us or by anyone. A section can be on screen and unread, and we will not invent the difference. | |
Every share the team owns, in one place. When somebody leaves, the links they sent keep working and their reading history stays with the workspace.